ARTICLE IN PRESS

Journal of Fluids and Structures 17 (2003) 927–939

www.elsevier.nl/locate/jnlabr/yjfls

Linear vibration analysis of cantilever plates partially

submerged in fluid
A. Ergin*, B. U&gurlu
Faculty of Naval Architecture and Ocean Engineering, Istanbul Technical University, Maslak, 80626, Istanbul, Turkey
Received 12 March 2002; accepted 19 March 2003

Abstract

Dynamic characteristics, such as natural frequencies and mode shapes, of cantilever plates, partially in contact with a
fluid, are investigated. In the analysis of the linear fluid–structure system, it is assumed that the fluid is ideal, and fluid
forces are associated with inertial effects of the surrounding fluid. This implies that the fluid pressure on the wetted
surface of the structure is in phase with the structural acceleration. Furthermore, the infinite frequency limit is assumed
on the free surface. The in vacuo dynamic properties of the plates are obtained by use of a standard finite-element
software. In the wet part of the analysis, it is assumed that the plate structure preserves its in vacuo mode shapes when
in contact with the surrounding fluid and that each mode shape gives rise to a corresponding surface pressure
distribution of the cantilever plate. The fluid–structure interaction effects are calculated in terms of the generalized
added-mass values independent of frequency (i.e., infinite frequency generalized added-masses), by use of a boundary-
integral equation method together with the method of images in order to impose the F = 0 boundary condition on the
free surface. To assess the influence of the surrounding fluid on the dynamic characteristics, the wet natural frequencies
and associated mode shapes were calculated, and they compared very well with the available experimental data and
numerical predictions.
r 2003 Elsevier Ltd. All rights reserved.

1. Introduction

An accurate understanding of the dynamic interaction between an elastic structure and fluid is necessary in various
engineering problems. Examples include vibration of floating structures (ships, offshore platforms, etc.) excited by wave
impact, water-retaining structures (dams, storage vessels, etc.) under earthquake loading, etc. The local resonant
vibration behavior of individual plates has also been a great concern to the shipbuilder and operator for many years.
Therefore, considerable effort has been made in the study of the vibrational response of plates, partially or totally
immersed in the fluid.
Lindholm et al. (1965) experimentally investigated the resonance frequencies of cantilever plates in air, totally or
partially immersed in water, and compared the measurements with the theoretical predictions. They evaluated the fluid
actions using a strip-theory approach. Meyerhoff (1970) calculated the added mass of thin rectangular plates in infinite
fluid, and described the potential flow around a flat rectangular plate by use of dipole singularities. The finite-element
method has also been applied to solve the fluid–structure interaction problems for completely submerged elastic plates
(see, for example, Muthuveerappan et al., 1979; Rao et al., 1993). On the other hand, Fu and Price (1987) studied the
dynamic behavior of a vertical or horizontal cantilever plate totally or partially immersed in fluid. In their analysis, they
calculated the generalized fluid loading to assess the influences of free surface and submerged plate length on the

*Corresponding author. Tel.:+90-212-285-6414; fax: +90-212-285-6454.

E-mail address: ergina@itu.edu.tr (A. Ergin).

0889-9746/03/$ - see front matter r 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0889-9746(03)00050-1
 ARTICLE IN PRESS
928 &
A. Ergin, B. Ugurlu / Journal of Fluids and Structures 17 (2003) 927–939

Fig. 1. Cantilever plate partially submerged in fluid.

dynamic characteristics. Recently, Liang et al. (2001) adopted an empirical added-mass formulation to determine the
frequencies and mode shapes of submerged cantilever plates, and compared their results with the available experimental
data and numerical predictions.
In this work, the dynamic characteristics (i.e., wet natural frequencies and mode shapes) of vibrating cantilever plates,
partially or totally immersed within the fluid, as illustrated in Fig. 1, are studied, assuming a linear fluid–structure
system. In this investigation, it is assumed that the fluid is ideal, i.e., inviscid and incompressible, and its motion is
irrotational. Furthermore, the fluid forces are associated with the inertial effect of the fluid. This means that the fluid
pressure on the wetted surface of the structure is in phase with the acceleration. In the analysis, it is assumed that the dry
plate vibrates in its in vacuo eigenmodes when it is in contact with fluid, and that each mode gives rise to a
corresponding surface pressure distribution on the wetted part of the structure. The in vacuo dynamic analysis entails
the vibration of the cantilever plate in the absence of any external force and structural damping, and the corresponding
dynamic characteristics (e.g., natural frequencies and principal mode shapes) of the plate structure were obtained using
ANSYS, a standard finite-element software.
At the fluid–structure interface, the body-boundary condition requires that the normal velocity of the fluid be equal
to that of the structure. The normal velocities on the wetted surface are expressed in terms of the modal structural
displacements, obtained from the in vacuo dynamic analysis. By use of a boundary-integral equation method in
conjunction with the method of images (i.e., imposing infinite frequency limit condition on the free surface), the fluid
pressure is eliminated from the problem and the fluid–structure interaction forces are calculated solely in terms of the
generalized added-mass coefficients that are independent of the vibrational frequency. As a result of the boundary
condition imposed on the free surface, the hydrodynamic damping becomes zero.
In this analysis, the wet surface is idealized by use of appropriate boundary elements, referred to as hydrodynamic
panels. The generalized structural mass matrix is merged with the generalized added-mass matrix and then the total
generalized mass matrix is used in solving the eigenvalue problem for the partially or totally immersed cantilever plate.
To assess the influence of the surrounding fluid on the dynamic behavior of the plate structure, wet natural frequencies
and associated mode shapes are calculated.
Comparison of the predicted dynamic characteristics with available experimental measurements (Lindholm et al.,
1965) and numerical calculations (Fu and Price, 1987) shows very good agreement. Furthermore, the influences of
various effects such as the submergence depth, plate aspect ratio (a/b) and thickness ratio (t/b) on the dynamic behavior
are also investigated.

2. Mathematical model

2.1. Formulation of the fluid problem

A right-handed Cartesian coordinate system, xyz, is adopted in the present study and it is shown in Fig. 1 for the
cantilever plate partially submerged. The coordinate system is fixed in space with its origin at O, and the x- and y-axis
coincide with the center-lines of the plate in the longitudinal and transverse directions, respectively. Meanwhile, the z-
axis lies across the plate thickness, and it is normal to the x–y plane.
 ARTICLE IN PRESS
&
A. Ergin, B. Ugurlu / Journal of Fluids and Structures 17 (2003) 927–939 929

Assuming an ideal fluid and irrotational motion, there exists a fluid velocity vector, v, defined as the gradient of the
velocity potential function F, where F satisfies Laplace’s equation throughout the fluid domain.
Before describing the responses of the flexible structure, it is necessary to assign coordinates to deflections at various
degrees of freedom. One particular set of generalized coordinates having significant advantage is the principal
coordinates of the dry structure (see, for example, Bishop et al., 1986; Ergin et al., 1992; Ergin, 1997a). For a structure
vibrating in an ideal fluid, with frequency o, the principal coordinate, describing the response of the structure in the rth
modal vibration, may be expressed by
pr ðtÞ ¼ pr eiot : ð1Þ
Therefore, the velocity potential function due to the distortion of the structure in the rth modal vibration may be
written as (see, e.g., Ergin and Temarel, 2002; Amabili et al., 1998)
Fr ðx; y; z; tÞ ¼ iofr ðx; y; zÞpr eiot ; r ¼ 1; 2; ::::; M; ð2Þ
where M represents the number of modes of interest.
On the wetted surface of the vibrating plate, the fluid normal velocity must be equal to the normal velocity of the
structure, and this condition for the rth modal vibration can be expressed as
@fr
 ¼ ur :n; ð3Þ
@n
where ur is the rth modal displacement vector of the median surface of the structure and n is the unit normal vector on
the wetted surface and it points into the fluid.
In this study, it is assumed that the plate structure vibrates at relatively high frequencies so that the effect of surface
waves can be neglected. Therefore, the free surface condition (infinite frequency limit condition) for fr can be
approximated by
fr ¼ 0; on the free surface: ð4Þ
The method of images (see, e.g., Kito, 1970; Ergin and Temarel, 2002) may be used, as shown in Fig. 2, to satisfy this
condition. By adding an imaginary boundary region, the condition given by Eq. (4) at the horizontal surface can be
omitted; thus the problem is reduced to a classical Neumann problem.
By use of Euler’s integral and neglecting the higher-order terms, the dynamic fluid pressure on the mean wetted-
surface of the flexible plate due to the rth modal vibration becomes
@Fr
Pr ðx; y; z; tÞ ¼ r : ð5Þ
@t
Substituting Eq. (2) into Eq. (5), the following expression for the pressure is obtained:
Pr ðx; y; z; tÞ ¼ ro2 fr ðx; y; zÞpr eiot : ð6Þ

Fig. 2. Wetted and imaginary surfaces for partially submerged plate.

 ARTICLE IN PRESS
930 &
A. Ergin, B. Ugurlu / Journal of Fluids and Structures 17 (2003) 927–939

The kth component of the generalized fluid–structure interaction force due to the rth modal vibration of the plate can
be expressed in terms of the pressure acting on the wetted surface of the plate as
Z Z Z Z
Zkr ðtÞ ¼ Pr uk n dS ¼ pr eiot ro2 fr uk n dS: ð7Þ
Sw Sw

The generalized added-mass term Akr can be defined as

Z Z
Akr ¼ r fr uk n dS: ð8Þ
Sw

Therefore, the generalized fluid–structure interaction force component, Zkr, can be rewritten as
Zkr ðtÞ ¼ Akr o2 pr eiot ¼ Akr p. r ðtÞ: ð9Þ
It should be noted, as a result of the infinite frequency limit condition imposed on the free surface, that the
generalized added-mass coefficients are constants and are independent of the vibrational frequency, and that the
hydrodynamic damping is zero.

2.2. Numerical evaluation of deformation potential f

The deformation potential, f, in a three-dimensional inviscid flow field due to the oscillating elastic structure can be
expressed by means of a distribution of unknown source strengths, s, over the wetted and imaginary surfaces of the
structure (see, e.g., Hess and Smith, 1967; Hess, 1975) in the following form:
Z Z
sðr0 Þ
fðrÞ ¼ dS; ð10Þ
Sw þSim Rðr; r0 Þ

where
 1
R ¼ ðx  x0 Þ2 þð y  y0 Þ2 þðz  z0 Þ2 2
and r = (x,y,z) denotes the position vector of the field point within the fluid, r0 ¼ ðx0 ; y0 ; z0 Þ is the position vector of the
source points on the wetted or image surface of the plate.
Substituting the boundary condition (3) into Eq. (10), the distribution of the unknown strengths si can be determined
from the set of following algebraic equations:
XN Z Z  
@ 1
2psi  sj dS ¼ uni ; i ¼ 1; 2; :::; N; ð11Þ
j¼1 DSj @n Rðrj ; rj Þ

where DSj represents the area of the jth panel, N is the number of panels used to discretize the wetted and imaginary
surfaces and uni denotes the modal displacement in the direction of the normal at the control point (xi, yi, zi) of the ith
panel.

2.3. Calculation of wet frequencies and mode shapes

It should be remembered that when a body oscillates in an unbounded, inviscid fluid, the generalized added-mass
coefficients are constants, independent of frequency, and the generalized fluid damping coefficients are zero. However,
in the case when the body oscillates in or near a free surface, the hydrodynamic coefficients exhibit frequency
dependence in the low-frequency region, but show a tendency towards a constant value in the high-frequency region. In
this study, it is assumed that the structure vibrates in the high-frequency region so that the generalized added-mass
values are constants and evaluated by use of Eq. (8). Furthermore, the generalized hydrodynamic damping and fluid
stiffness are assumed to be zero. Hence, the generalized equation of motion for the dynamic fluid–structure interaction
system (see, e.g., Ergin, 1997b, c), assuming free vibrations with no structural damping, is
 
o2 ða þ AÞ þ c p ¼ 0 ð12Þ
where a and c denote the generalized structural mass and stiffness matrices, respectively. The matrix A represents the
infinite frequency generalized added-mass coefficients.
Solving the eigenvalue problem, expressed by Eq. (12), yields the wet frequencies and associated mode shapes of the
cantilever plate in contact with the fluid. To each wet frequency or, there is a corresponding wet eigenvector por ¼
fpr1 ; pr2;:::; prm g satisfying Eq. (12). The corresponding uncoupled wet mode shapes for the structure partially or totally in
 ARTICLE IN PRESS
&
A. Ergin, B. Ugurlu / Journal of Fluids and Structures 17 (2003) 927–939 931

contact with fluid are obtained as

X
M
u% r ðx; y; zÞ ¼ fu% r ; v%r ; w% r g ¼ uj ðx; y; zÞprj ; ð13Þ
j¼1

where uj ðx; y; zÞ ¼ fuj ; vj ; wj g denote the in vacuo mode shapes of the dry cantilever plate and M the number of mode
shapes included in the analysis. It should be noted that the fluid–structure interaction forces associated with the inertial
effect of the fluid do not have the same spatial distribution as those of the in vacuo modal forms. Consequently, this
produces hydrodynamic coupling between the in vacuo modes. This coupling effect is introduced into Eq. (12) through
the generalized added-mass matrix A.

3. Numerical results and comparisons

A series of calculations have been performed in order to demonstrate the applicability of the aforementioned theory
to vibrating cantilever plates, partially or totally immersed within the fluid. Two cantilever plates with different aspect
and thickness ratios were considered. The first plate chosen has length a = 1016 mm, width b = 203.2 mm, thickness t
= 4.84 mm (see Fig. 1), and it was experimentally studied by Lindholm et al. (1965). The second plate adopted in the
present calculations is 10 m long, 10 m wide, 0.238 m thick, and it was numerically investigated by Fu and Price (1987)
by use of a three-dimensional hydroelasticity theory. Both plates are made of steel and have the following material
characteristics: Young’s modulus = 206.8 GPa, Poisson’s ratio = 0.3, and mass density = 7830 kg/m3. Fresh water is
used as the surrounding fluid with a density of 1000 kg/m3.

3.1. Idealization and convergence

The in vacuo dynamic characteristics of cantilever plates were obtained by use of ANSYS (1994), a finite-element
software. This produces information on the natural frequencies and principal mode shapes of the dry structure. In these
calculations, the plates were discretized by four-noded, quadrilateral shell elements, including both membrane and
bending stiffness influences.
In a preliminary calculation, 128 elements were distributed over the cantilever plate of Lindholm et al. (1965). The
distribution over the plate consists of 16 and 8 equally spaced elements, respectively, along the length and width of the
plate structure. To test the convergence of the calculated dynamic properties (natural frequencies and principal mode
shapes), the number of elements over the cantilever plate was increased first to 288 (24 elements along the length and 12
elements along the width of the plate) and then to 512 (32 elements along the length and 16 elements along the width of
the cantilever). Table 1 shows the experimental results of Lindholm et al. (1965) and calculated natural frequencies
obtained from ANSYS, for the first six modes. The differences in the results indicate that the calculated values are
converging with increasing number of elements. The results of the final idealization (512 elements) compare very well
with the experimental measurements of Lindholm et al. (1965); hence they were adopted for the in vacuo dynamic
properties of the cantilever plate. It should be noted that the mode shapes are divided into two distinct groups: the
symmetric (S) (bending) and antisymmetric (A) (torsional) distortional mode shapes. A symmetric mode shape shows
similar displacement characteristics on either side of the Ox-axis (see Fig. 1), and, on the other hand, an antisymmetric
mode shape exhibits opposite displacement behavior about the Ox-axis.
Another series of calculations were performed in order to test the convergence of the hydrodynamic properties
(generalized added-mass terms). The aim of this exercise was to represent accurately the distortional mode shapes of the

Table 1
Convergence of FEM (in vacuo) natural frequencies (Hz)

Mode Mode type 128 elements 288 elements 512 elements Experiment(Lindholm et al.,1965)

1 S 3.94 3.94 3.94 3.84

2 S 24.65 24.66 24.66 24.20
3 A 38.70 38.99 39.07 39.10
4 S 69.42 69.28 69.24 68.10
5 A 118.17 119.17 119.47 121.00
6 S 136.49 136.64 136.35 —
 932 &
A. Ergin, B. Ugurlu

Table 2
Convergence of wet natural frequencies (Hz)

Mode Mode type Depth ratio (d/a)

0.25 0.5 0.75 1

96 Panels 192 Panels 320 Panels 176 Panels 360 Panels 608 Panels 256 Panels 528 Panels 896 Panels 336 Panels 696 Panels 1184 Panels

1 S 2.21 2.26 2.29 1.85 1.89 1.90 1.78 1.81 1.83 1.78 1.81 1.82
2 S 21.07 21.23 21.32 15.29 15.56 15.69 11.99 12.13 12.20 11.51 11.62 11.68
3 A 28.56 29.17 29.42 24.72 25.23 25.43 23.52 23.98 24.16 23.37 23.83 24.00
4 S 58.63 59.01 59.32 52.16 52.13 52.23 38.39 38.52 38.67 34.35 34.28 34.31
ARTICLE IN PRESS

5 A 103.42 104.57 105.01 91.41 93.08 93.84 75.28 76.70 77.31 72.96 73.40 73.91
6 S 117.97 116.84 116.66 100.11 98.91 98.71 85.49 84.55 84.46 72.19 71.58 71.22
/ Journal of Fluids and Structures 17 (2003) 927–939
 ARTICLE IN PRESS
&
A. Ergin, B. Ugurlu / Journal of Fluids and Structures 17 (2003) 927–939 933

Table 3
Comparisons of dry and wet natural frequencies (Hz)

Mode Mode type This study Experiment (Lindholm et al. 1965)

In vacuo Depth ratio (d/a) In -air Depth ratio (d/a)

0.25 0.5 0.75 1 0.25 0.5 0.75 1

1 S 3.94 2.29 1.90 1.83 1.82 3.84 2.17 1.82 1.79 1.78
2 S 24.66 21.32 15.69 12.20 11.68 24.20 21.01 15.5 11.99 11.50
3 A 39.07 29.42 25.43 24.16 24.00 39.10 29.75 25.50 24.20 24.20
4 S 69.24 59.32 52.23 38.67 34.31 68.10 57.36 51.61 38.27 33.50
5 A 119.47 105.01 93.84 77.31 73.91 121.00 106.35 95.99 79.00 75.26
6 S 136.35 116.64 98.71 84.46 71.22 — — — — —

wetted surface of the cantilever plate. For the plate of Lindholm et al. (1965), the tests were carried out separately for
four different depth ratios, namely d/a = 0.25, 0.5, 0.75 and 1. A test case may simply be identified by the number of
hydrodynamic panels distributed along the length (NPL) and along the width of the wetted plate. However, the
hydrodynamic panels are distributed over the wetted top and bottom plate surfaces separated by the thickness of the
plate and over the wetted edge surfaces parallel to the x–z or y–z planes (see Fig. 1). In the first group of idealizations,
96 (NPL = 4), 176 (NPL = 8), 256 (NPL = 12) and 336 (NPL = 16) quadratic hydrodynamic panels (boundary
elements) were distributed over the wetted surfaces of the cantilever plate, respectively, for the depth ratios d/a = 0.25,
0.5, 0.75 and 1. The number of panels along the width of the wetted plate was eight in the first group of idealizations.
Subsequently, the number of panels over the wetted surfaces was increased to 192 (NPL = 6), 360 (NPL = 12), 528
(NPL = 18) and 696 (NPL = 24), and finally to 320 (NPL = 8), 608 (NPL = 16), 896 (NPL = 24) and 1184 (NPL =
32) for d/a = 0.25, 0.5, 0.75 and 1, respectively. The number of hydrodynamic panels along the width was 12 and 16 for
the second and final group of idealizations, respectively. Table 2 shows the convergence of the predicted wet natural
frequencies with increasing number of hydrodynamic panels for the first six wet modes. The differences between the
results, based on the last two groups of idealizations, are reasonably small for all the submergence depths. Therefore, it
may be said that the final group of idealizations (i.e., 320, 608, 896 and 1184 panel idealizations) adequately represents
the distortional shapes of the partially submerged cantilever plate, and they were adopted for the calculations presented
in Tables 3 and 4, and for those in Figs. 4 and 5. An additional convergence study was also carried out to establish the
number of distortional modes needed for the predictions. As a result of this analysis, 12 in vacuo modes were included
in the calculations.

3.2. Calculated results and comparisons

By solving the eigenvalue problem, Eq. (12), the uncoupled modes and associated frequencies of the cantilever plate
of Lindholm et al. (1965), partially in contact with fluid, were obtained. The calculated natural frequencies are
compared with the experimental measurements of Lindholm et al. (1965) in Table 3 for the plate oscillating in air (or in
vacuo) or partially submerged in water (i.e., d/a = 0.25, 0.5, 0.75 and 1). Although, there are some differences between
the predicted and measured results, the predicted wet natural frequencies show very good agreement with the
corresponding experimental data. The differences lie in the range between 0.2% and 5.5%. For the same plate, Fig. 3
shows the in vacuo principal mode shapes (finite element results) whereas the calculated wet modes are presented in Fig.
4 for the depth ratio d/a = 0.5. The mode shapes shown in Figs. 3 and 4, corresponding to the first six wet (or dry)
natural frequencies, are simply ordered with increasing frequency. As can be realized from the comparison of Fig. 4
with Fig. 3, the wet mode shapes vary slightly from the in vacuo modes. This is the hydrodynamic coupling of the in
vacuo modes which creates negligible influence on the wet distortional mode shapes when the plate is partially
submerged. Similar distortional mode shapes were also obtained for the depth ratios d/a = 0,25, 0.75 and 1, and
therefore they are not presented here. It can also be observed from Table 3 that the frequencies behave as expected.
That is to say the frequencies decrease with increasing area of contact with fluid. The largest area of contact was
in the case of the fully submerged cantilever plate (d/a = 1). Therefore, the lowest frequencies occurred in this case
(see Table 3).
Table 4 shows the calculated generalized added-mass terms for the depth ratios d/a = 0.25, 0.5, 0.75 and 1. The
generalized added-mass terms in Table 4 are presented for the first 12 distortional in vacuo modes and they correspond
 ARTICLE IN PRESS
934 &
A. Ergin, B. Ugurlu / Journal of Fluids and Structures 17 (2003) 927–939

Fig. 3. In vacuo mode shapes of cantilever plate (FE results): (a) first mode, frequency 3.94 Hz; (b) second mode, frequency 24.66 Hz;
(c) third mode, frequency 39.07 Hz; (d) fourth mode, frequency 69.24 Hz; (e) fifth mode, frequency 119.47 Hz; (f) sixth mode, frequency
136.35 Hz.

Fig. 4. Wet mode shapes of cantilever plate for depth ratio d/a = 0.5: (a) first mode, frequency 1.9 Hz; (b) second mode, frequency
15.69 Hz; (c) third mode, frequency 25.43 Hz; (d) fourth mode, frequency 52.23 Hz; (e) fifth mode, frequency 93.84 Hz; (f) sixth mode
frequency 98.71 Hz.

to a generalized structural mass of 1 kg m2. Here, it is assumed that the structure preserves its in vacuo principal mode
shapes in the fluid and that each mode gives rise to the surface pressure distribution of the flexible structure. However,
the hydrodynamic forces associated with the inertial effect of the surrounding medium do not necessarily have the same
spatial distribution as those of the in vacuo principal modes. Consequently this produces hydrodynamic coupling
between the in vacuo modes. From Table 4, it is seen that the generalized added-mass matrices are symmetric and the
 ARTICLE IN PRESS
&
A. Ergin, B. Ugurlu / Journal of Fluids and Structures 17 (2003) 927–939 935

Table 4
Generalised added-mass coefficients (kg m2) of partially submerged cantilever plate

Mode 1 2 3 4 5 6 7 8 9 10 11 12

d/a Mode Mode type S S A S A S S A S A S A

0.25 1 S 1.956 1.034 0.002 0.325 0.001 0.182 0.000 0.001 0.411 0.003 0.391 0.002
2 S 1.034 0.693 0.000 0.400 0.000 0.152 0.000 0.000 0.019 0.000 0.109 0.000
3 A 0.002 0.000 0.738 0.000 0.587 0.000 0.000 0.333 0.000 0.064 0.000 0.124
4 S 0.325 0.400 0.000 0.428 0.000 0.383 0.000 0.001 0.270 0.001 0.115 0.001
5 A 0.001 0.000 0.587 0.000 0.492 0.000 0.000 0.324 0.000 0.134 0.000 0.017
6 S 0.182 0.152 0.000 0.383 0.000 0.498 0.000 0.001 0.451 0.002 0.286 0.001
7 S 0.000 0.000 0.000 0.000 0.000 0.000 0.005 0.000 0.000 0.000 0.000 0.000
8 A 0.001 0.000 0.333 0.001 0.324 0.001 0.000 0.296 0.000 0.237 0.000 0.153
9 S 0.411 0.019 0.000 0.270 0.000 0.451 0.000 0.000 0.485 0.001 0.372 0.001
10 A 0.003 0.000 0.064 0.001 0.134 0.002 0.000 0.237 0.001 0.313 0.000 0.307
11 S 0.391 0.109 0.000 0.115 0.000 0.286 0.000 0.000 0.372 0.000 0.368 0.000
12 A 0.002 0.000 0.124 0.001 0.017 0.001 0.000 0.153 0.001 0.307 0.000 0.372

0.5 1 S 3.301 0.316 0.001 0.719 0.001 0.440 0.000 0.001 0.023 0.001 0.002 0.001
2 S 0.316 1.424 0.002 1.019 0.002 0.086 0.000 0.001 0.581 0.002 0.213 0.001
3 A 0.001 0.002 1.349 0.000 0.472 0.000 0.000 0.231 0.000 0.193 0.000 0.071
4 S 0.719 1.019 0.000 1.337 0.001 0.602 0.000 0.001 0.178 0.001 0.324 0.000
5 A 0.001 0.002 0.472 0.001 0.624 0.000 0.000 0.480 0.000 0.030 0.000 0.214
6 S 0.440 0.086 0.000 0.602 0.000 1.074 0.000 0.001 0.707 0.001 0.070 0.001
7 S 0.000 0.000 0.000 0.000 0.000 0.000 0.008 0.000 0.000 0.000 0.000 0.000
8 A 0.001 0.001 0.231 0.001 0.480 0.001 0.000 0.862 0.000 0.428 0.000 0.125
9 S 0.023 0.581 0.000 0.178 0.000 0.707 0.000 0.000 1.059 0.001 0.524 0.000
10 A 0.001 0.002 0.193 0.001 0.030 0.001 0.000 0.428 0.001 0.667 0.000 0.431
11 S 0.002 0.213 0.000 0.324 0.000 0.070 0.000 0.000 0.524 0.000 0.874 0.000
12 A 0.001 0.001 0.071 0.000 0.214 0.001 0.000 0.125 0.000 0.431 0.000 0.721

0.75 1 S 3.660 0.418 0.000 0.447 0.001 0.108 0.000 0.001 0.211 0.000 0.079 0.000
2 S 0.418 3.117 0.001 0.149 0.002 0.702 0.000 0.002 0.106 0.001 0.213 0.000
3 A 0.000 0.001 1.614 0.000 0.043 0.000 0.000 0.165 0.000 0.063 0.000 0.110
4 S 0.447 0.149 0.000 2.203 0.002 0.560 0.000 0.002 0.729 0.002 0.093 0.000
5 A 0.001 0.002 0.043 0.002 1.366 0.001 0.000 0.268 0.000 0.337 0.000 0.158
6 S 0.108 0.702 0.000 0.560 0.001 1.746 0.000 0.001 0.510 0.001 0.525 0.000
7 S 0.000 0.000 0.000 0.000 0.000 0.000 0.008 0.000 0.000 0.000 0.000 0.000
8 S 0.001 0.002 0.165 0.002 0.268 0.001 0.000 1.126 0.000 0.354 0.000 0.335
9 S 0.211 0.106 0.000 0.729 0.000 0.510 0.000 0.000 1.603 0.000 0.355 0.000
10 A 0.000 0.001 0.063 0.002 0.337 0.001 0.000 0.354 0.000 1.045 0.000 0.289
11 S 0.079 0.213 0.000 0.093 0.000 0.525 0.000 0.000 0.355 0.000 1.453 0.000
12 A 0.000 0.000 0.110 0.000 0.158 0.000 0.000 0.335 0.000 0.289 0.000 1.038

1 1 S 3.699 0.534 0.000 0.288 0.000 0.242 0.000 0.000 0.133 0.000 0.103 0.000
2 S 0.534 3.515 0.000 0.435 0.000 0.176 0.000 0.000 0.229 0.000 0.085 0.000
3 A 0.000 0.000 1.649 0.000 0.050 0.000 0.000 0.046 0.000 0.046 0.000 0.038
4 S 0.288 0.435 0.000 3.134 0.000 0.340 0.000 0.000 0.097 0.000 0.207 0.000
5 A 0.000 0.000 0.050 0.000 1.613 0.000 0.000 0.051 0.000 0.040 0.000 0.046
6 S 0.242 0.176 0.000 0.340 0.000 2.719 0.000 0.000 0.266 0.000 0.060 0.000
7 S 0.000 0.000 0.000 0.000 0.000 0.000 0.008 0.000 0.000 0.000 0.000 0.000
8 A 0.000 0.000 0.046 0.000 0.051 0.000 0.000 1.549 0.000 0.053 0.000 0.035
9 S 0.133 0.229 0.000 0.097 0.000 0.266 0.000 0.000 2.348 0.000 0.213 0.000
10 A 0.000 0.000 0.046 0.000 0.040 0.000 0.000 0.053 0.000 1.466 0.000 0.055
11 S 0.103 0.085 0.000 0.207 0.000 0.060 0.000 0.000 0.213 0.000 2.028 0.000
12 A 0.000 0.000 0.038 0.000 0.046 0.000 0.000 0.035 0.000 0.055 0.000 1.371
 ARTICLE IN PRESS
936 &
A. Ergin, B. Ugurlu / Journal of Fluids and Structures 17 (2003) 927–939

cross-coupling terms are small in comparison with the diagonal ones. On the other hand, it can also be observed from
Table 4 that the coupling effect generally becomes stronger for small submergence depths. For instance, the cross-
coupling terms are mainly larger for the depth ratio, d/a, of 0.25 when compared with the fully submerged plate (depth
ratio d/a = 1). Furthermore, it is also observed from Table 4 that there is negligibly small coupling between the
symmetric (bending) and antisymmetric (torsional) modes.

Fig. 5. Variation of wet frequencies with aspect ratio, a/b, and thickness ratio, t/b, for depth ratio d/a = 0.5: (a) first mode; (b) second
mode; (c) third mode; (d) fourth mode; (e) fifth mode (f) sixth mode—, t/b = 0.0238;– – –, t/b = 0.0611;–   –   –,t/b = 0.124.
 ARTICLE IN PRESS
&
A. Ergin, B. Ugurlu / Journal of Fluids and Structures 17 (2003) 927–939 937

Fig. 5 (continued).

The analysis was subsequently extended to investigate the effects of plate aspect ratio, a/b, and thickness ratio, t/b, on
the dynamic characteristics of this partially submerged cantilever plate. Fig. 5 shows the wet natural frequencies of the
cantilever plate for three different thickness ratios, t/b, (i.e., 0.0238, 0.0611 and 0.124) and five different aspect ratios, a/
b, (i.e., 1, 2, 3, 4 and 5), for the first six wet modes and depth ratio d/a =0.5. As can be observed from Fig. 5, the wet
natural frequencies decrease with increasing a/b values. However, they increase with increasing t/b values.
 ARTICLE IN PRESS
938 &
A. Ergin, B. Ugurlu / Journal of Fluids and Structures 17 (2003) 927–939

Table 5
Comparisons of dry and wet natural frequencies (Hz)

Mode Mode type This study Fu and Price (1987)

In vacuo Depth ratio (d/a) In vacuo Depth ratio (d/a)

0.25 0.5 0.75 1 0.25 0.5 0.75 1

1 S 2.04 1.58 1.30 1.20 1.17 2.06 1.63 1.31 1.21 1.17
2 A 4.98 4.09 3.51 3.31 3.28 5.05 4.20 3.48 3.26 3.22
3 S 12.48 11.46 10.23 8.60 8.00 12.70 11.66 10.41 8.70 8.03
4 S 15.82 13.66 12.06 11.22 11.07 16.10 13.86 12.20 11.38 11.21
5 A 18.04 16.61 15.37 13.21 12.58 18.40 16.81 15.62 13.25 12.55
6 S 31.23 29.08 27.39 23.96 22.98 — — — — —

In a further study, another cantilever plate, studied numerically by Fu and Price (1987), was adopted in the
calculations. The in vacuo dynamic characteristics (i.e., natural frequencies and mode shapes) of the plate were obtained
by using 256 finite elements (16 equally spaced elements along the length and width of the plate structure). The
calculated natural frequencies are presented and compared with the predications of Fu and Price (1987) in Table 5. In
their analysis, Fu and Price (1987) adopted 64 four-node thin plate elements (eight elements along the length and width
of the plate). All the wet natural frequencies in Table 5 were evaluated by use of Eq. (12). The values of the current
method compare well with those obtained from the theory adopted by Fu and Price (1987), as can be seen in Table 5.
However, there are some differences between the results of the current method and those of Fu and Price (1987). The
maximum difference is 4.5% and was observed for the depth ratio d/a = 0.25. For the converged results of the current
method presented in Table 5, the wet surface of the cantilever plate is idealized by using 176, 320, 464, and 608 panels,
respectively, for the depth ratios d/a =0.25, 0.5, 0.75 and 1. The number of panels distributed along the length of the
plate was 4, 8, 12, and 16, respectively, for the depth ratios d/a = 0.25, 0.50, 0.75 and 1. In all cases, 16 panels were
distributed along the width of the plate. On the other hand, 12 in vacuo modes were adopted for the converged results
of the current method presented in Table 5.

4. Conclusions

The dynamic characteristics (wet natural frequencies and associated modes) of two different, partially submerged
cantilever plates were calculated by an approach based on the boundary-integral equation method and the method of
images. It can be concluded from the results presented here that the method proposed is suitable for relatively high-
frequency vibrations of partially submerged elastic structures.
From the results presented, the calculations based on the present method show very good agreement with the
experimental data of Lindholm et al. (1965). The differences lie within the limits one would expect when
comparing experimental data with numerical predictions. On the other hand, the results of the present method are
also compared with the predictions of Fu and Price (1987) in Table 5, and there is good agreement between the two
calculations. The predictions based on the current method are slightly lower in comparison with those of Fu and Price
(1987).
As can be seen from Table 4, the generalized added-mass matrices are symmetric, and off-diagonal terms represent
the effect of coupling between the in vacuo modes. It can also be concluded from Table 4 that the coupling becomes
stronger with decreasing submergence depths.
It can also be concluded from the results in Tables 3 and 5 that the wet frequencies behave as expected from the
theory. That is to say that the wet frequencies decrease with increasing submergence depth. Therefore, the lowest wet
frequencies occur in the cases of fully submerged cantilever plates.
As can be seen from Fig. 5, the wet natural frequencies decrease with increasing aspect ratio, a/b, but increase with
increasing thickness ratio, t/b.
The present study has demonstrated the versatility of the method through two different cantilever plates vibrating
partially or fully submerged in water. In a future study, the effect of a flowing fluid on the dynamic characteristics of
elastic structures will be considered.
 ARTICLE IN PRESS
&
A. Ergin, B. Ugurlu / Journal of Fluids and Structures 17 (2003) 927–939 939

Acknowledgements

This research was financially supported by Scientific and Technical Research Council of Turkey (Project No:
ICTAG-1842). This support is gratefully acknowledged.

References

Amabili, M., Pa.ıdoussis, M.P., Lakis, A.A., 1998. Vibrations of partially filled cylindrical tanks with ring-stiffeners and flexible
bottom. Journal of Sound and Vibration 213, 259–299.
ANSYS. 1994. User’s Manual. ANSYS, Inc., Houston.
Bishop, R.E.D., Price, W.G., Wu, Y., 1986. A general linear hydroelasticity theory of floating structures moving in a seaway.
Philosophical Transactions of the Royal Society of London A 316, 375–426.
Ergin, A., 1997a. The response behavior of a submerged cylindrical shells using the doubly asymptotic approximation method (DAA).
Computers and Structures 62, 1025–1034.
Ergin, A., 1997b. An approximate method for the free vibration analysis of partially filled and submerged, horizontal cylindrical shells.
Journal of Sound and Vibration 207, 761–767.
Ergin, A., 1997c. Dynamic characteristics of partially filled and submerged cylindrical tanks. Proceedings of the Eighth International
Congress on Marine Technology, IMAM, Vol.2, Istanbul, pp 21–30.
Ergin, A., Price, W.G., Randall, R., Temarel, P., 1992. Dynamic characteristics of a submerged, flexible cylinder vibrating in finite
water depths. Journal of Ship Research 36, 154–167.
Ergin, A., Temarel, P., 2002. Free vibration of a partially liquid-filled and submerged, horizontal cylindrical shell. Journal of Sound
and Vibration 254, 951–965.
Fu, Y., Price, W.G., 1987. Interactions between a partially or totally immersed vibrating cantilever plate and the surrounding fluid.
Journal of Sound and Vibration 118, 495–513.
Hess, J.L., 1975. Review of integral-equation techniques for solving potential-flow problems with emphasis on the surface-source
method. Computer Methods in Applied Mechanics and Engineering 5, 145–196.
Hess, J.L., Smith, A.M.O., 1967. Calculation of potential flow about arbitrary bodies. Kuchemann, D., et al.,(Eds.), Progress in
Aeronautical Sciences. Vol. 8, pp.1–138.
Kito, F., 1970. Principles of Hydro-elasticity. Keio University Publication, Tokyo.
Liang, C.-C., Liao, C.-C, Tai, Y.-S., Lai, W.-H., 2001. The free vibration analysis of submerged cantilever plates. Ocean Engineering
28, 1225–1245.
Lindholm, U.S, Kana, D.D., Chu, W.-H., Abramson, H.N, 1965. Elastic vibration characteristics of cantilever plates in water. Journal
of Ship Research 9, 11–22.
Meyerhoff, W.K., 1970. Added masses of thin rectangular plates calculated from potential theory. Journal of Ship Research 14,
100–111.
Muthuveerappan, G., Ganesan, N., Veluswami, M.A., 1979. A note on vibration of a cantilever plate immersed in water. Journal of
Sound and Vibration 63, 385–391.
Rao, P.S., Sinha, G., Mukhopadhyay, M., 1993. Vibration of submerged stiffened plates by the finite element method. International
Shipbuilding Progress 40, 261–292.